Schizophrenia (SCZ) is a disabling neurodevelopmental disorder causing profound cognitive impairment. SCZ is hypothesized to arise from synaptic disturbances affecting large-scale neural connectivity. This view is supported by neuroimaging studies that repeatedly show alterations in prefrontal cortex (PFC) function and connectivity and disruptions across thalamo-cortical and associative cortex circuits. However, the complex neurobiology of early-course SCZ remains uncharacterized, limiting treatments for early illness phases when intervention is crucial. This is a major objective for improving targeted therapies, predicting prognosis, and promoting early detection. Our overarching goal is to longitudinally characterize concurrent functional and structural dysconnectivity in early-course SCZ in relation to cognitive deficits via state-of-the-art neuroimaging. In turn, we aim to inform synaptic hypotheses underlying clinical neuroimaging effects via biophysically-based computational modeling scaled to the level of neural networks. To address these knowledge gaps, we will examine longitudinal progression of neural dysconnectivity in early-course SCZ patients after their initial admission into the Specialized Treatment Early in Psychosis (STEP) Clinic at Yale. In turn, we will follow patients longitudinally at 6, 12, and 24 months later in comparison with 50 matched healthy controls. To quantify dysconnectivity the project will use leading functional and structural methods optimized by the Human Connectome Project (HCP), in line with the NIMH Connectomes Related to Human Disease initiative. First, we aim to test if the recently identified PFC and thalamo-cortical markers exhibit concurrent (or dissociable) structural and functional alterations. This balanced longitudinal design can distinguish `state' versus `trait' neuroimaging markers during early illness course in relation to clinically-relevant variables. Specifically, examining effects of pharmacotherapy, treatment compliance, duration of untreated psychosis, and symptom severity, informs the clinical utility of these promising neuroimaging markers. Second, the project will test if these neuroimaging markers relate to severity of cognitive deficits - a hallmak clinical feature of SCZ. We aim to concurrently examine working memory (WM) via our validated neuroimaging paradigms to test if specific aspects of structural and functional dysconnectivity predict WM deficits. This provides a much-needed link between dysconnectivity and cognitive impairment in SCZ. Finally, to inform synaptic hypotheses behind neural dysconnectivity, such as cortical excitation-inhibition (E/I) imbalance resulting from hypo-function of the N-methyl-D-aspartate glutamate receptor (NMDAR), we aim to use biophysically-based computational models that incorporate relevant cellular detail. We aim to iteratively explore synaptic parameters governing E/I balance by fitting in silico effects with in vivo clinical neuroimaging findings. This computational psychiatry approach can help interpret dynamic neural dysconnectivity in SCZ via computational fits and yield new synaptic targets for treatment studies focused on early SCZ stages, when intervention is most vital.